Skip to content

周伟奇 / bmw-ocr

Go to a project
Toggle navigation
Toggle navigation pinning
24977d55 by 冯轩
Options

ADD:财报逻辑

1 parent 5c3c9d17
Inline Side-by-side
Showing 4 changed files with 218 additions and 10 deletions
src/apps/doc/consts.py
......@@ -1247,6 +1247,7 @@ FSM_CONTRACT_WEP_FIELD = 'fsm_wep_ocr'
FSM_CONTRACT_MSI_FIELD = 'fsm_msi_ocr'
FSM_CONTRACT_SC_FIELD = 'fsm_sc_ocr'
FSM_CONTRACT_SC2_FIELD = 'fsm_sc2_ocr'
FS_FIELD = 'fs_ocr'
BS_CLASSIFY = 10089
......@@ -1504,6 +1505,7 @@ SE_BD_FIELD = ['被保险人姓名', '被保险人证件号码', '车架号', '
JDMPV_VALUE = ['-', '--', '0%', '0.00', '/', '0', '']
JYPZ_TYPE_1 = ['二手车交易凭证']
JYPZ_TYPE_2 = ['BMW官方认证二手车交易凭证', '二手车交易凭证']
SE_FS_FIELD = ['Hash值', '公章', '财年', '资产负债表内容', '利润表内容']
SE_BANK_FIELD = ['accountNo', 'bankName']
SE_DDA_FIELD = ['applicationId(1)', 'applicationId(2)', 'bankName', 'companyName', 'customerName', 'idNum', 'accountHolderName', 'accountNo']
......@@ -1817,8 +1819,11 @@ BS_EN = 'Bank Statement'
HIL_CONTRACT_1_EN = '售后回租合同'
HIL_CONTRACT_2_EN = '车辆租赁抵押合同'
HIL_CONTRACT_3_EN = '车辆处置协议'
FS_EN = 'Financial Statement'
FSS_EN = 'Financial Statement Supplementary'
DDA_NO_FIND = '需人工查看DDA或截图'
FS_NO_FIND = '未提供财报或财报不完整'
SKIP_CARD = {SME_BL_EN}
......@@ -2105,16 +2110,6 @@ HT_QRS_COMPARE_LOGIC = {
'合同编号': ('合同编号', 'se_qrs_compare', {}, '请确认是否提供合格的送达地址确认书'),
}
FINANCIAL_STATEMENT_COMPARE_LOGIC = {
'Hash值': ('Hash值', '', {}, '财报Hash值与系统不一致'),
'公章': ('公章', '', {}, '财报无公章'),
}
FINANCIAL_STATEMENT_SUPPLEMENTARY_COMPARE_LOGIC = {
'公司名称': ('公司名称', '', {}, '财报情况说明公司名称错误'),
'公章': ('公章', '', {}, '财报情况说明无公章'),
}
HT_COMPARE_LOGIC = {
'合同编号-每页': ('合同编号-每页', 'se_list_compare', {}, '合同编号与系统不一致'),
# '合同编号-每页(no-asp)': ('合同编号-每页', 'se_list_compare', {'pop_last': True}, '合同编号与系统不一致'),
......@@ -2169,6 +2164,15 @@ HT_COMPARE_LOGIC = {
'无ASP产品': ('无ASP产品', 'se_self_compare_other_asp', {}, GZS_REASON_1),
}
# 格式：'excel字段名':('数据库字段名','比对逻辑','特殊处理可以忽略','比对不合格时的返回内容')
FS_COMPARE_LOGIC = {
'Hash值': ('code', 'hash_code_compare', {}, '财报Hash值与系统不一致'),
'公章': ('stamp', 'stamp_dict_compare', {}, '财报无公章'),
'财年': ('财年', 'fiscal_year_compare', {}, '财报所属财年错误'),
'资产负债表内容': ('资产负债表内容', 'ocr_list_not_zero_compare', {}, '财报资产负债表为空'),
'利润表内容': ('利润表内容', 'ocr_list_not_zero_compare', {}, '财报利润表为空'),
}
AFC_HT_DATE_FIELDS = ['主借人日期', '共借人日期', '保证人日期1', '保证人日期2']
# MVC_OCR_FIELD = 'mvc_ocr'
......@@ -2197,6 +2201,7 @@ SE_COMPARE_FIELD = {
HIL_CONTRACT_1_EN: (HIL_CONTRACT_1_FIELD, HIL_CONTRACT_1_COMPARE_LOGIC, False),
HIL_CONTRACT_2_EN: (HIL_CONTRACT_2_FIELD, HIL_CONTRACT_2_COMPARE_LOGIC, False),
HIL_CONTRACT_3_EN: (HIL_CONTRACT_3_FIELD, HIL_CONTRACT_3_COMPARE_LOGIC, False),
FS_EN: (FS_FIELD, FS_COMPARE_LOGIC, False),
}
SE_COMPARE_FIELD_AUTO = {
......
src/celery_compare/tasks.py
......@@ -953,6 +953,12 @@ def get_se_cms_compare_info_auto(application_id, last_obj, application_entity, d
if individual_info['applicantType'] == consts.APPLICANT_TYPE_ORDER[0] and not is_corporate:
is_cdfl_bo = True
# CHINARPA-4660 是否公户判断
is_bo_tccor = False
customersubType = individual_info.get('customersubType', '')
if individual_info['applicantType'] == consts.APPLICANT_TYPE_ORDER[0] and customersubType == 'TCCOR':
is_bo_tccor = True
for id_info in individual_info.get('IDInformation', []):
if id_info.get('idType') in consts.SE_CMS_FIRST_ID_FIELD_MAPPING:
license_en, is_prc = consts.SE_CMS_FIRST_ID_FIELD_MAPPING[id_info['idType']]
......@@ -1120,6 +1126,24 @@ def get_se_cms_compare_info_auto(application_id, last_obj, application_entity, d
# vehicle_info[consts.UCI_EN] = vehicle_field_input
compare_info['vehicleInfo'] = vehicle_info
# 公户财务报表-------------------------------------------------------------------------------------------------
financial_statement_info = {}
financial_statement_input = []
hashCode = cms_info.get('hashCode', '')
fiscalYear = cms_info.get('fiscalYear', '')
totalAssets = cms_info.get('totalAssets', 0)
totalLiabilitiesAndOwnersEquity = cms_info.get('totalLiabilitiesAndOwnersEquity', 0)
netProfit = cms_info.get('netProfit', 0)
if is_bo_tccor:
financial_statement_input.append((consts.SE_FS_FIELD[0], hashCode))
financial_statement_input.append((consts.SE_FS_FIELD[1], consts.SE_STAMP_VALUE))
financial_statement_input.append((consts.SE_FS_FIELD[2], fiscalYear))
financial_statement_input.append((consts.SE_FS_FIELD[3], [totalAssets, totalLiabilitiesAndOwnersEquity]))
financial_statement_input.append((consts.SE_FS_FIELD[4], netProfit))
financial_statement_info[consts.FS_EN] = financial_statement_input
compare_info['financialStatementInfo'] = financial_statement_info
# 银行卡-------------------------------------------------------------------------------------------------------
bank_info = {}
bank_name = cms_info.get('bankAccountDetails', {}).get('bankName', '')
......@@ -2523,6 +2547,107 @@ def se_compare_license(license_en, ocr_res_dict, field_list):
return result_field_list, no_ocr_result, field_img_path_dict
def se_fs_compare(license_en, ocr_res_dict, field_list):
ocr_field, compare_logic, special_expiry_date = consts.SE_COMPARE_FIELD[license_en]
is_find = False
no_ocr_result = False
special_expiry_date_slice = False
result_field_list = []
section_img_info = dict()
field_img_path_dict = dict()
ocr_res_str = ocr_res_dict.get(ocr_field)
if ocr_res_str is not None:
ocr_res_list = json.loads(ocr_res_str)
length = len(ocr_res_list)
# 先判断最后一次上传的文件是不是包括3个，如果不是直接返回"未提供财报或财报不完整"
last_ocr_str = ocr_res_list[length-1]
if len(last_ocr_str.get('code',{})) != 3 or len(last_ocr_str.get('stamp',{})) != 3:
compare_log.info('{0} [se_fs_compare] last ocr result len < 3'.format(log_base))
else:
for res_idx in range(length-1, -1, -1):
if is_find:
break
for idx, (name, value) in enumerate(field_list):
ocr_str = ocr_res_list[res_idx].get(compare_logic[name][0])
#hash值/公章 不需要ocr结果，所以即使为空也可以进入比对
if isinstance(ocr_str, str) or ocr_str == 'code' or ocr_str == 'stamp':
result = getattr(cp, compare_logic[name][1])(value, ocr_str, **compare_logic[name][2])
no_key = False
else:
result = consts.RESULT_N
ocr_str = empty_str
no_key = True
if idx == 0 and result == consts.RESULT_N and length > 1:
break
is_find = True
section_img_info[consts.SECTION_IMG_PATH_KEY] = ocr_res_list[res_idx].get(consts.SECTION_IMG_PATH_KEY, '')
section_img_info[consts.ALL_POSITION_KEY] = ocr_res_list[res_idx].get(consts.ALL_POSITION_KEY, {})
if special_expiry_date:
section_img_info[consts.SECTION_IMG_PATH_KEY_2] = ocr_res_list[res_idx].get(
consts.SECTION_IMG_PATH_KEY_2, '')
section_img_info[consts.ALL_POSITION_KEY_2] = ocr_res_list[res_idx].get(consts.ALL_POSITION_KEY_2, {})
img_path = ocr_res_list[res_idx].get(consts.IMG_PATH_KEY, '') if result == consts.RESULT_N else empty_str
if isinstance(value, list):
value = json.dumps(value, ensure_ascii=False)
error_type = empty_error_type if result == consts.RESULT_Y else ErrorType.OCR.value
result_field_list.append((name, value, result, ocr_str, img_path, error_type, compare_logic[name][3]))
else:
no_ocr_result = True
if not is_find:
for name, value in field_list:
if isinstance(value, list):
value = json.dumps(value, ensure_ascii=False)
no_find_str = consts.FS_NO_FIND
result_field_list.append((name, value, consts.RESULT_N, empty_str, empty_str, ErrorType.NF.value, no_find_str))
if is_find:
section_img_path = section_img_info.get(consts.SECTION_IMG_PATH_KEY, '')
if os.path.exists(section_img_path):
failed_field = []
base_img_path = empty_str
for name, _, result, _, img_path, _, _ in result_field_list:
if result == consts.RESULT_N:
if special_expiry_date_slice and name == 'idExpiryDate':
continue
failed_field.append(name)
if base_img_path == empty_str:
base_img_path = img_path
if len(failed_field) > 0:
info = section_img_info.get(consts.ALL_POSITION_KEY, {})
section_position = info.get(consts.POSITION_KEY, {})
section_angle = info.get(consts.ANGLE_KEY, 0)
try:
last_img = img_process(section_img_path, section_position, section_angle)
except Exception as e:
for field in failed_field:
field_img_path_dict[field] = base_img_path
else:
pre, suf = os.path.splitext(section_img_path)
for field in failed_field:
try:
res_field = compare_logic[field][0]
is_valid, coord_tuple = field_build_coordinates(info.get(res_field, {}))
if is_valid:
save_path = '{0}_{1}{2}'.format(pre, field, suf)
field_img = last_img[coord_tuple[0]:coord_tuple[1], coord_tuple[2]:coord_tuple[3], :]
cv2.imwrite(save_path, field_img)
field_img_path_dict[field] = save_path
else:
field_img_path_dict[field] = base_img_path
except Exception as e:
field_img_path_dict[field] = base_img_path
return result_field_list, field_img_path_dict
def se_compare_license_id(license_en, id_res_list, field_list, is_auto):
if is_auto:
......@@ -3023,6 +3148,7 @@ def se_compare_process(compare_info, ocr_res_dict, is_gsyh, is_auto, id_res_list
rpa_failure_reason = {}
field_result_dict = {}
# compare_info 格式： {'financialStatementInfo': {'Financial Statement': [('11', '22'), ('111', '222'), ('1111', '2222')]}}
for info_key, info_value in compare_info.items():
if info_key in ['individualCusInfo', 'applicantInformation']:
for idx, license_list in info_value.items():
......@@ -3104,6 +3230,8 @@ def se_compare_process(compare_info, ocr_res_dict, is_gsyh, is_auto, id_res_list
result_field_list, field_img_path_dict = se_contract_qrs_compare(license_en, ocr_res_dict, strip_list)
elif license_en == consts.BS_EN:
result_field_list, field_img_path_dict = se_bs_compare(license_en, ocr_res_dict, strip_list, is_auto, aa_type)
elif license_en == consts.FS_EN:
result_field_list, field_img_path_dict = se_fs_compare(license_en, ocr_res_dict, strip_list, is_auto, aa_type)
else:
result_field_list, _, field_img_path_dict = se_compare_license(license_en, ocr_res_dict, strip_list)
......
src/common/tools/comparison.py
......@@ -676,6 +676,50 @@ class Comparison:
except Exception as e:
return self.RESULT_N
def hash_code_compare(self, input_str, ocr_dict, **kwargs):
try:
balance_sheet_hash = ocr_dict.get('balance_sheet','')
income_statement_hash = ocr_dict.get('income_statement','')
cash_flow_statement_hash = ocr_dict.get('cash_flow_statement','')
if balance_sheet_hash != input_str or income_statement_hash != input_str or cash_flow_statement_hash != input_str:
return self.RESULT_N
else:
return self.RESULT_Y
except Exception as e:
return self.RESULT_N
def stamp_dict_compare(self, input_str, ocr_dict, **kwargs):
try:
balance_sheet_stamp = ocr_dict.get('balance_sheet','')
income_statement_stamp = ocr_dict.get('income_statement','')
cash_flow_statement_stamp = ocr_dict.get('cash_flow_statement','')
if balance_sheet_stamp != 1 or income_statement_stamp != 1 or cash_flow_statement_stamp != 1:
return self.RESULT_N
else:
return self.RESULT_Y
except Exception as e:
return self.RESULT_N
def fiscal_year_compare(self, input_str, ocr_list, **kwargs):
try:
this_year_str = datetime.now().strftime('%Y')
this_year = int(this_year_str)
last_year = this_year - 1
if input_str != this_year and input_str != last_year:
return self.RESULT_N
return self.RESULT_Y
except Exception as e:
return self.RESULT_N
def input_list_not_zero_compare(self, input_list, ocr_list, **kwargs):
try:
for item in input_list:
if item == 0:
return self.RESULT_N
return self.RESULT_Y
except Exception as e:
return self.RESULT_N
cp = Comparison()
......
src/common/tools/mssql_script28.py 0 → 100644
import pyodbc
hil_sql = """
ALTER TABLE hil_ocr_result ADD fs_ocr nvarchar(max);
ALTER TABLE hil_se_ocr_result ADD fs_ocr nvarchar(max);
ALTER TABLE hil_ocr_result ADD fss_ocr nvarchar(max);
ALTER TABLE hil_se_ocr_result ADD fss_ocr nvarchar(max);
"""
afc_sql = """
ALTER TABLE afc_ocr_result ADD fs_ocr nvarchar(max);
ALTER TABLE afc_se_ocr_result ADD fs_ocr nvarchar(max);
ALTER TABLE afc_ocr_result ADD fss_ocr nvarchar(max);
ALTER TABLE afc_se_ocr_result ADD fss_ocr nvarchar(max);
"""
hil_cnxn = pyodbc.connect('DRIVER={ODBC Driver 17 for SQL Server};', autocommit=True)
hil_cursor = hil_cnxn.cursor()
hil_cursor.execute(hil_sql)
hil_cursor.close()
hil_cnxn.close()
afc_cnxn = pyodbc.connect('DRIVER={ODBC Driver 17 for SQL Server};', autocommit=True)
afc_cursor = afc_cnxn.cursor()
afc_cursor.execute(afc_sql)
afc_cursor.close()
afc_cnxn.close()
Styling with Markdown is supported
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!